2860 MARTIN : RESEARCHES ON SILICON COMPOUNDS. PART VII.CCLXVI1.-Researches on Silicon Compounds. PartThe Action of Ethyl Alcohol on Disilicon VII.Hexachloride.By GEOFFREY MARTIN.WHEN ethyl alcohol acts on disilicon hexachloride, Si2Cl,, thefollowing nine substances are, theoretically, capable of being pro-duced, none of which has hitherto been described:SiCl, SiCI, SiCl,*OEt Sic 1, SiCl,*OEtI ISiCI,*OEt dia(OEt), \ SiCI,*OEt J &i(OEt), L &Cl(OEt), 2One form. Two forms. Two forms. Y YSi(OEt), SiCl(OEt), Si(OEt), Si(OEt),ISiCl,*OEt diCl(OEt), diCl(OEt), di(OEt),Two forms. One form. One form.MARTIN : RESEARCHES ON SILICON COMPOUNDS. PART VII. 2861The production of these substances has also considerabletheoretical importance, because the chlorine atoms containedtherein can be easily replaced by hydroxy-groups merely by treat-ment' with water, and the properties of the resulting hydroxy-com-pounds have some interest', as they afford means of verifying theauthor's theory* (Be?-., 1912, 45, 2097; 1913, 46, 3289) thatdirectly united silicon atoms, in the presence of attached oxygenatoms, are decomposed by alkalis with the evolution of hydrogen,according t o the scheme:III -51-1Thus each direct Si-Si linking corresponds with the evolution ofone molecule of evolved hydrogen.The properties and descrip-tion of these hydroxy-compounds are best left t o a laher paper.The present paper is confined t o the preparation and propertiesof the above-mentioned chlorinated compounds.Having prepared a considerable amount of pure disilicon hexa-chloride as a starting point, the author was able t o prepare in apure condition compounds of the following formulze : Si,Cl,*OEt,Si,C"l,(OEt),, Si,Cl,(OEt),, Si,Cl(OEt),, and Si,(OEt),.It will be noticed that' each of the substances Si,Cl,(OEt),,Si,Cl,(OEt),, and Si,Cl,(OEt), can, theoretically, exist - in twostructurally different forms.However, in actual practice theauthor up to the present was unable to find more than one modifi-cation of each of the& forms. Either the t'wo modifications ofeach variety boil a t the same temperature, or, what is more likely,the action proceeds almost t o completion in one direction only, theone modification being produced in overwhelming amount and theother in traces, so that i t is tlifficult t o isolate the two isomerides.One curious fact that appears when alcohol is added t o disiliconhexachloride is the circumstance that, apparently, a most vigorousaction sets in, the mixture appearing to boil with the copiousevolution of hydrogen chloride. Nevertheless, such an intense coldis produced t h a t hoar-frost collects on the sides of the flask.* I n view of Kipping's footnote (this vol., p.484) I wish to draw attention to thepaper (T,, 1913, 103, 119) in which I have dealt in detail with the accusations nowrepeated by him. With Kipt'ing's suggestion that readers should refer t o andcoinpare the original papers bearing on the matter a t issue I am in full agreement.Iu regard t o Kipping's statement that my preliminary note (Bey., 1912, 45,2097) was published without his permission, I would poiiit out that as theexperiment^, were desigued and carried out by myself aloire, Kipping's pernissionwas not necessary.VOL. cv.9 2862 MARTIN : RESEARCHES ON SILICON COMPOUNDS. PART VII.This is an example of a vigorous chemical action taking placewith the absorption of so much heat that water can easily be frozenthereby.The feebleness of the chemical forces tending t o bring about theinterchange of the chlorine atoms of the disilicon hexachloride forethoxy-groups is also, no doubt, responsible for the fact that whendisilicon hexachloride is treated with excess of alcohol, substitutionof ethoxy-groups for chlorine does not proceed quantitatively withthe production of the compound Si,(OEt),, thus :Si,Cl, + 6EtOH = Si,(OEt), + 6HC1,but a mixture of lower chlorides is formed, and the last chlorineatoms of the disilicon hexachloride are quite difficult to displaceby ethoxy-groups, repeated heating and distillation with excess ofalcohol being necessary befosre the last traces of chlorine are ex-pelled and the compound Si,(OEt), is obtained.It is true that the first equivalents of chlorine of the disiliconhsxachloride molecule are readily displaced by ethoxy-groups, yetas each successive chlorine atom is removed the displacement pro-ceeds with greater and greater difficulty, so that the substancesSi,Cl,(OEt), and Si,Cl(OEt), do not any longer fume very notice-ably in air, and are scarcely acted on by ethyl alcohol a t theordinary temperature.They must be heated with alcohol to. loooand above before any visible action, such as the evolution ofhydrogen chloride, takes place.The separation of these chlorinated silicon compounds by frac-tional distillation in a pure state proved a most difficult task, veryprolonged and numerous fractionations being necessary beforecomplete separation of the various components was effected. Thedifficulties were enormously increased by the fact that these sub-stancm react with traces of moisture, giving white, explosive pre-cipitates, and causing a turbidity i n the resulting liquids. Con-sequently, every vessel used in these repeated fractionations hadto be most carefully freed from every trace of moisture by carefullywashing out with alcohol, followed by ether, heating almost t oredness in a luminous gas flame while a current of dry air waspassed through the flask, and affixing fresh from the heating andwhile still fairly hot to the fractionating apparatus.Only thuscan absolutely clear, colourless fluids, free from every trace ofturbidity, be obtained. Moreover, in the fractionating apparatuscorks had to be dispensed with as far as possible, and f o r this reasonthe long fractionating columns had to be fused on to the flasks,the union by means of a cork proving unsatisfactory.Rubber corks, after a time, were attacked by the liquids, be-coming hard and crackedMARTIN : RESEARCHES ON SILICON COMPOUNDS.PART VII. 2863Other difficulties also arose, on account of the corrosive actionof these liquids on the skin. They set up painful sores which tooka long time to heal, and as in the repeated fractionations the changing of the vessels made contact of the fingers with t'hese liquidsalmost unavoidable (owing to traces adhering to the sides, etc., ofthe vessels), i t was found highly advisable to protect the thumband fingers with indiarubber coverings.I n the following table are compared the physical properties ofthe different members of the, series:Substance.Si,Cl,( OEt)Si,Cl,( OEt),Si,Cl,( OEt),Si,Cl,( OEt),Si,Cl(OEt),Sizc16si2( OEt),Boiling pointunder 34 mm. Density.60.5" 1 * 5 624:"84 1 * 3 8 8:'104 1 27 Ojo122 1.163y138 1-092i7141 0.97 1 8:7- -Refractiveindex(D line).1.4748 at 18"1.4688 ,, 14.5"1.4432 ,, 14.5"1.4333 y y 14.5"1.4205 y y 14.5"1.4134 ,, 14.5"-It will be seen that as we proceed down the series from Si,Cl,to Si,(OEt),, there is a progressive increase in the value of theboiling points; the substitution of a single chlorine atom in Si,CI,for an ethoxy-group causes an increase in the boiling point of23*5O/34 rnm., but the successive displacements of chlorine byethoxyl effeds a rapidly diminishing value in the rise of the boil-ing point, until finally the last two members of the series, namely,Si,Cl(OEt), and Si,(OEt),, boil a t nearly the same temperature-so near together, in fact, that it is difficult t o separate a mixtureof these two substances by fractional distillation.The same gradation of physical properties is apparent when thedensities are compared.Thus, whilst disilicon hexachloride has adensity of 1 *56, hexaethoxysilico-ethare, Si,(OEt), (D 0*97), isactually lighter than water ; the intermediate members of the serieshave intermediate values.EXPERIMENTAL.Some preliminary experiments were first, made. In one case7 grams of disilicon hexachlaride were treated with about 100 C.C.of ethyl alcohol (99.8 per cent'.), and in another case 21 grams ofdisilicon hexachloride were treated with 50 C.C. of ethyl alcohol.In each case a vigorous action took place, much hydrogenchloride was evolved, and the liquid became so cold that the flaskscontaining t'he liquid became covered with ice, whilst a thermo-meter placed in one flask registered - go.Two layers of liquid wereobserved t o form.After distilling over the excess of alcohol the residual oils decom-9 a 2864 MARTIN : RESEARCEES ON SILICON COMPOUNDS. PART VII.posed on distillation under the ordinary pressures, the thermometerrising rapidly t o 200°, and dense brown fumes appeared in the flaskand a brown mass remained behind.Under diminished pressure, however, colourless oils distilled overwithout decomposition, but they were evidently mixtures, sinceunder 15 mm. pressure some liquid distilled at, 65-96O, but about75 per cent. passed over at 110-115°/15 mm., after which thethermometer rose and a small quantity of liquid distilled over a t170-180a/ 15 mm.Attempts t o fractionate these liquids showed that they consistedof compIex mixtures of different substances, and that when disiliconhexachloride is treated with excess of alcohol the action does notproceed quantitatively, thus : Si,Cl, + 6EtOH = Si,(OEt), + 6HC1,but leads to the production of intermediate chlorinated products invery considerable quantity.All these colourless oils cmtained chlorine', and when thrown intowater they produced white precipitates, which dissolved on warming(but not in the cold) with sodium or potassium hydroxides with theevolution of hydrogen. They also evolved hydrogen with ammonia.When thrown into water, however, the change, from an oil t o awhite solid took some little time, aiid no heat perceptibls to thehand (although perceptible t o a thermometer) was evolved.When exposed to the air in a dish the oils became transformedinto transparent, solid glassss, which also possessed the power ofdissolving in warm potassium hydroxide' with thel evolution ofhydrogen.That it was atmospheric moisiure (and not oxygen) that con-verted t'lie oils into glasses on exposure to air was proved by thefact that when equal qiantities of the oils were exposed (1) to dryair in a desiccator (over sulphuric acid), and (2) to the ordinarymoist air outside the desiccator, i t was found that, only the surfaceexposed to the moist air outside the desiccator exhibited this solidi-fying effect.These preliminary experiments proved that the action of ethylalcohol on disilicon hexachloride was no simple one, and that inorder to isolate pure' products considerable quantities of disiliconhexachloride would have to be treated with alcohol, and the result-ing liquid very carefully fractionated.Accordirgly, in all, some 600 grams of disilicon hexachloridewere treated with ethyl alcohol, and the products isolated aftermany months' fractional distillation, as described belowMARTIN : RESEARCHES ON SILICON COMPOUNDS. PART VII.2865Pen tuc hloro e t hoxysilico-e t hane , Si,Cl,=OEt, and T e trachloro-diethoxysilico-ethane, Si,Cl,(OEt),.Eighty-four grams of disilicon hexachloride (1 mol.) were placedin a flask fitted with a reflux Condenser, and 15 grams (1.05 mol.)of ethyl alcohol were gradually run in (with continual shaking)through a stoppered funnel.A vigorous action took place as the alcohol entered, much hydro-gen chloride was evolved, and the liquid in the flask becameslightly yellow and very cold, so that ice was deposited on the sidesof the containing flask.A few drops of an orange-coloured liquidappeared to float’ on the surface of the liquid in the flask.On heating, however, under a reflug condenser on the water-bath for one hour these orange-coloured drops gradually dissp-peared, the liquid finally appearing almost colourless. After leavingovernight fhe liquid was distilled under diminished pressure, usinga Young’s 15-rod-and-disk fractionating column fused on t o thedistilling flask.However, a great many fractionations proved that the fraction-ating column used was not efficient enough t o separate sharply thevarious components of the mixture, and i t was also evident thatlarger quantities of material would have t o be employed in orderto obtain pure products. Four main fractions were isolated:The liquid boiled fairly con-s€antly a t 60-65O/35 mm., then for some time a t 75O/35 mm.Thethermometer rose slowly t o 80°/35 mm.The bulk passed over a t S2--S7°,/35 mm.Fraction 1.-Up to S0°/35 mm.Fraction 2.-80-90°/35 mm.Fraction 3.-90-100°/35 mm.Fraction 4.-100-110°/35 mm.The experiment was repeated three times, using the followingquantities : (I) 101 grams of disilicon hexachloride and 21 grams ofethyl alcohol. (3) 105grams and 22 grams respectively.This makes in all about 390 gramsof disilicon hexachloride treated with 79 grams of ethyl alcohol.Each of these fractions was then fractionated, using a Young’s20-rod-and-disk fractionating column, with the result that therewere finally obtained, after a very prolonged series of fractiona-tions :(1) A colourless, fuming liquid, boiling a t 50-5Z0/35 mm. and131-136O/ 767 mm., which contained silicon and chlorine#, but wasnot disilicon hexachloride (which boils a t 145O/760 mm.). The yieldwas, however, only 8 c.c., and it was certainly not pure, so that i twas not further examined.(2) 100 grams and 21 grams respectively2866 MARTIN : RESEARCHES ON SILICON COMPOUNDS. PART VII.(2) About 17 C.C. of a fuming, colourless liquid, boiling a t59-61°/35 mm.and 144--146O/767 mm., which was unchangeddisilicon hexachloride,(3) One hundred grams of a colourlem, mobile, fuming liquid,boiling a t 83'5-84*5O/34 nim., which was pentachloroethoxysilico-ethane, Si,Cl,*OEt (see below).(4) Fifty-four grams of a colourless, fuming liquid, boiling a t1 0 4 O / 35 mm., which proved to be tetrachlorodiethoxysilico-ethane.(5) About 155 grams of residues separated from the variousfractionations in isolating the fractions (3) and (4).Since both the liquids (3) and (4) boiled constantly, and allefforts to alter their boiling points proved unavailing, they wereanalysed as follows :The liquid was introduced into a weighing bottle (without touch-ing the sides, etc., of the latter) by means of a m a l l pipette fittedwith a rubber teat, the stopper quickly inserted, and the wholeweighed.Wa€,er was then quickly introduced into the weighingbottle in order t o decompose the chloride. The white solid whichappeared was firs€ €rested with concentrated ammonia (whichcaused the evolution of hydrogen, and thus loosened the precipitateadhering t o the sides of the vessd), and the liquid and precipitatewere rinsed out into a beaker and warmed for some time on thewater-bath with a little ammoniz until the effervescence of hydrogenceased. The liquid could be then either acidified with nitric acidand the chlorine estimated by the Volhard method, or exactlyneutralised with nitric acid, and the chlorine estimated by titratingwith _V/ 10-silver nitrate, using potassium chromate as indicator.I n the present case the clear liquid was filtered from the precipi-tate, the latter being'well washed, and the liquid was neutralisedwith nitric acid, made LIP t o 250 c.c., and titrated with N/IO-silvernitrate.Liquid (3) (b. p.83*5-84'5°/35 mm.) :0.5910 gave C1= 63.7.The liquid, therefore, was practically pure pentachloroethoxy-silico-ethane, Si,Cl,*OEt.Liquid (4) (b. p. 10P0/35 mm.):0.8990 gave Cl=49.7.This liquid, therefore, needed some pmrification.C,H,OCl,Si, requires C1= 63.6 per cent.C,H,,O,Cl,Si, requires C1= 49.1 per cent.It was treatedwith a very small amount oE alcohol to remove the excess of chlorine,and the liquid was again fractionated, the first and last portionsof the distillate being rejected.The liquid finally obtained, whichboiled a t the same temperature as that given above, was found tMARTIN : RESEARCHES ON SILICON COMPOUNDS. PAW VII. 2867contain C1= 49.2, and so was practically pure tetrachlorodiethoxy-silico-ethane, Si,Cl,(OEt),.Pentachtloroethoxysilico-ct hane, Si,Cl,-OEG, is a mobile, colourless,fuming liquid, boiling a t 83*5--84.5O/35 mm., having D? 1-388and nF5 1.4568. It does not solidify when immersed in a freezingmixture of ice and salt. When exposed t o moisture i t is convertedinto a white solid, which can be made t.0 explode by touching witha hot glass rod or even by brushing with a test-tube brush. A fulldescription of these explosive hydroxy-derivatives is reeerved for alater paper.Tetrachlorodiethozysilico-ethane, Si,Cl,(OEt),, is a mobile, colour-less, fuming liquid, boiling a t 104O/34 mm., and having Dia 1.270and 7 ~ : ’ ~ 1.4432.It does not freeze when immersed in a mixtureof ice and salt. It is decomposed by water to form an explosivewhite hydroxy-compound.The preceding results prove that in general when one equivalentof alcohol acts on disilicon hexachloride the reaction by no meansproceeds quantitatively, thus :SkC1, + EtOH = Si,Cl,*OEt + HCI.Although pentachloroethoxysilico-ethane is th0 main product, avery considerable amount (more than 30 per cent.) of tetrachloro-diethoxysilico-ethane is produced a t the same time, and the conse-quence is that some free disilicon hexachloride is left uncombined,and can actually be separated from the mixture by fractionaldistillation.Trichlorot riet htoxysilico-e t hane, Si,Cl,(OEt),.From the previous distillations there had accumulated about155 grams of residues, consisting mainly of a mixture of penta-chloroethoxy- and tetrachlorodiethoxy-silico-ethane, together withsome trichlorotriethoxysilico-ethane and similar product&. I n orderto convert this mixture into the trichloro-derivative, 55 C.C. of ethylalcohol (99.8 per cent.) wer0 gradually added and the mixture con-stantly shaken, and heated on the water-bath for one hour.Asthe ethyl alcohol entered a vigorous action ensued, much hydrogenchloride was evolved, and the flask became very cold. After leavingovernight the liquid was fractionated under diminished pressure,using a Young’s 20-rod-and-disk fractionating column fused on tothe distilling flask.At115-1 1 6 O / 35 mm.the thermometer remained constant until about10-15 C.C. of li2uid had distilled over. The bulk of the liquidpassed over between 123O and 125O’/35 mm., about 80 C.C. beinghere collected, and the thermometer then rose to 130--135O/35 mm.Only a few C.C. passed over between 60° and 115O/35 mm2868 MARTIN : RESEARCHES ON SILICON COMPOUNDS. PART VII.On cooling, the residuei in the flask solidified after standing over-night, forming a dirty grey mass, which melted when gentlyheated.The various fractions were now subjected t o a prolonged andcareful series of fractionations, whereby there was isolated about50 grams of a liquid boiling constantly a t 123--124O/35 mm.:0.5741 gave C1= 35.6.C,H,,0,C1,Si2 requires C1= 35.7 per cent.Trichlorotriethoxysilico-etha?ze is a colourless, fuming liquid,boiling a t 123--124O/35 mm., and having D;O 1.163 and Y I : ' ~1.4333.It' does not solidify when immersed in a freezing mixtureof ice1 and salt. It is soon decomposed by atmospheric moisture,yielding a white, explosive' hydroxy-compound. The liquid actsstrongly on the1 skin, causing deep and painful wounds, which donot readily heal.A t t empt to Isolate Dichloro t etrae thoxysiZico-efl~a?ze, Si2C1,(OEt),.From the previous distillations there had accumulated about11 1 grams of liquid, largely consisting of trichlorotriethoxysilico-ethane, together with some tetrachlorodietlioxy- and dichlorotetra-ethoxy-silico-ethane.This mixture was placed in a flask, and about60 C.C. of ethyl alcohol were run in with constant shaking. Muchhydrogen chloride was evolved, and the liquid became very cold.The mixture was then subjected t o fractional distillation. Nearlyall passed over between 130° and 137O/33 mm., the bulk distillinga t 131--132O/33 nim.After prolonged fractionation a colourless liquid was isolated,boiling a t 132--133O/35 mm., 121°/23 mm., 116O/18 mm., and110°/12 mm.Nearly 70 grams of this liquid were isolated, and this was shownby analysis (Found, C1=17.3) to be a mixture of dichlorotetra-ethoxy- (Cl = 23.0) and chloropentaethoxy-silico-ethane (C1= 11.2).The mixtme was subjected t o prolonged fractional distillation,whereby some of the first fractions had their chlorine contentraised to 18.7 per cent.However, i t was not found possible toobtain dichlorotetraethoxysilico-ethane, Si,Cl,(OEt),, in a purecondition, the boiling points of the mono- and di-chloro-derivativesbeing so close together as t o make a complete separation of both byfractional distillation alone a matter of considerable difficulty, a tleast with the small quantities available.Chlo ropentae t hox ysilico-e t hnn P , Si,C1 (OE t)5.With the object of isolating the above compound, some of theresidues and fractions from previous distillations, amounting in alMARTIN : RESEARCHES ON STLICON COMPOUNDS. PART VII. 2869t o about 66 grams, were collected and placed in a flask, and about9 grams of ethyl alcohol were added.On adding the alcohol, how-ever, apparently no action took place a t the ordinary temperature,the liquid becoming neither sensibly hot nor cold to the hand, norwas any visible amount of hydrogen chloride evolved. On heat-ing, however, an action set in, some hydrogen chloride beingevolved.It is thus evident that as the substitution of the chlorine forethoxy-groups in the disilicon hexachloride molecule proceeds, theaction takes place less and less readily, substances like dichloro-tetraethoxysilico-ethane not being noticeably acted on by alcohola t the ordinary t'emperature, although this action takes placereadily enough on heating.During the distillation, however, the flask, while being heatedon a metal-bath, burst, and the liquid was lost.A fresh attempt was therefore made to isolate chloropentaet'hoxy-silicoiethane by acting on disilicon hexachloride with the theoreticalamount of alcohol.Disilicon liexachloride (138 grams; 1 mol.) was placed in a flaskand treated with 99.8 ethyl alcohol (120 grams).Hydrogenchloride was evolved, and the flask became cold. The liquid,after being heat'ed on the water-bath t o complete the action, wasfractionally distilled under diminished pressure, the heating beingcarried out on a metal-bath. After about 80 grams of a liquidboiling a t 120--124O/15 mm. had distilled over, amd while theflask was nearly half full of residual liquid, a violent explosiontook place, the flask being shattered, and the fragments hurledvert8dcally into the air with such force that an indentation was madein the hard plaster on the ceiling, 4 o r 5 metres above the workingbench.The employment of a metal-bath, therefore, for heating the1 flaskappeared inadvisable, as the explosion possibly arose from over-heating the residues of high boiling point left in the flask.Aswill be shown in a subsequent paper, many compounds containingsilicon atoms directly united are explosive under certain conditions.In all the subsequent distillations the flasks were heated on oil-baths, the temperatures of which were carefully controlled so asto avoid overheating, and with this precaution no further explosionswere met with in the course of many subsequent distillations.The prece'ding experimentl was once more repeated, with precau-tions against overheating.Disilicon hexachloride (1 38 grams) wasplaced in a flask, and 99.8 per cent. ethyl alcohol (120 grams) wasgradually added. Hydrogen chloride was evolved, and cold wasproduced. The liquid was then heated on the water-bath for on28'10 MARTIN : RESEARCHES ON SILICON COMPOUNDS. PART VII.hour until the evolution of hydrogen chloride had ceased, and theflask and its contents were allowed to remain overnight.The weight of the contents of the flask after the evolution ofhydrogen chloride had ceased was 167 grams, so that 138 gramsof disilicon hexachloride had yielded about 167 grams of productand lost 91 grams of hydrogen chloride.According to the equation Si,Cl, + 5EtOH = Si,Cl(OEt), + 5IIC1,138 grams of disilicon hexachloride should give 161 grams ofchloropentaethoxysilico-ethane and lose 93 grams of hydrogenchloride.The numbers actually found agree sufficiently closely with thetheoretical to make certain that the action had taken place almostquantitatively.The liquid was therefore fractiofnally distilleld, the fractionatingflask being immersed in an oil-bath, the temperature of which wasnot allowed to rise above 175O in order to avoid the danger ofoverheating.After about 20 grams of liquid had passed over between 75Oand 120°/15 mm., the main fraction of 112 grams passed over,boiling very constantly, between 120° and 123O/15 mm.I n theflask remained about 25 grams of residue of higher boiling point.The main fraction, boiling at 120-123°/15 mm., was now sub-jected to fradioaal distillation, using a Young's 20-rod-and-diskcolumn fused on t o the distilling flask, which was heated in an oil-bath t'o about 1 6 0 O .There was thus obhined a fraction boilingvery constantly at about 121°/15 mm.However, the substance was certainly not quite pure, a chlorinedetermination, carried out as previously described, using chromatea6 indicator, giving C1=12*0 (Calc., Cl=ll.2 per cent.). Con-sequently, the excess of chlorine was removed by adding a verylittle alcohol t o the liqu*id, and then subjecting it to fractional andrepeated distillation, neglecting the first and last parts of thedistillate.There wits thus obtained the, main bulk of liquid, which, whenthree successive fractions were taken, all boiled a t the sametemperature of 126--127O/19 mm., and seemed quite pure.A chlorine estimation was next carried out with each of thethree successive fractions, and gave, respectively, C1= 11-5,There is no doubt, therefore, that the substance was fairly purechloropentaethoxysilico-ethane, which requires C1= 11-2 per cent.Chloropentae t hoxysilico-e thane, Si,Cl(OE t)5, is a colourlessliquid (D;' 1.092, ng'5 1.4205) which does not fume noticeably inthe air, but on exposure t o atmospheric moisture gives a white,c1= 11 '2, c1= 11.1MARTIN : RESEARCHES ON SILICON COMPOUNDS.PART VII. 2871amorphous, explosive hydroxy-compound. The liquid does notsolidify when immersed in a freezing mixture of ice and salt.Theboiling points were determiried under different pressures, asfollowsi :Pressure ............... 13 mm. 14mm. 15mm. 16 mm. 18 mm. 20 mm. 22 mm.Boilingpoint ......... 119" 120" 121' 122" 124' 127' 129OPressure ............... 23 mm. 24 mm. 27 mm. 30 mm. 32 mm. 36 mm.Boilingpoint ......... 130" 131' 134" 136" 137" 139'Eexae t hoxysilico-e thane, Si,(OEt),.About 151 grams of chloropentaethoxysilico-ethane, prepared aspreviously described, and boiling a t about 120-121°/15 mm., wasplaced in a flask fitted with a reflux condenser. On adding about32 C.C. of ethyl alcohol (1 mol.), no visible action appeared to takeplace. On heating on the oil-bath, however, the evolution ofhydrogen chloride began a t about looo, and became quite vigoromwhen the temperature of the oil-bath reached 1 10-120°, showingthat the action, Si2C1(OEt), + EtOH= Si,(OEt), + HC1, was pro-ceeding.On distilling the liquid, which passed over for the most partbetween 120° and 124O/15 mm., no alcohol distilled over, so thatt'he absorption of alcohol was complete.However, a little chlorinewas present in the distillate, and in order to remove this about30 C.C. of alcohol were added, and the liquid was heated on thewater-bath under a reflux condenser. The excess of alcohol couldbe distinctly seen distilling up the condenser, condensing, anddropping back on the liquid in the flask, so that excess of alcoholwas certainly present.The alco,hol was distilled over, and in i t was found a perceptibleamount, of chlorine. The liquid was therefore repeatedly boiledwith alcohol, the excess of alcohol being distilled off each time. Itwas found very difficult, even when excess of alcohol was repeatedlypresent, to expel every trace of chlorine, proving that the abovereaction by no means readily proceeds to completion.However, there was finally obtained a colourless, non-fumingliquid boiling a t 123O/ 15 mm., which, after repeated fractionationwith a Young 20-rod-and-disk fractionating column fused on to thedistilling flask, was found to be free from chlorine. In analysingthe liquid, the silicon was estimated by heating some of the sub-stance in a platinum crucible with concentrated sulphuric acid,igniting, and weighing the resulting silica :0-2800 gave 0.4499 CO, and 0.2354 H,O. C=43-8; H=9-4;1.1281 gave 0.4120 SiO,.Si = 17.2.C,,H,,O,Si, requires C= 44-07 ; H = 9.2 ; Si = 17.3 per cent2572 BRADY AND DUNN :Hexaethoxysilico-etJ~ane is a colourless, non-fuming oil whichdoes not solidify in a freezing mixture of ice and salt. It boils a t123*/15 mm., 132*/24 mm., 137*5O/30 mm., and 141°/34 mm.,and has D:' 0.9718 and 12:'~ 1.4134. The boiling points are thusvery close t o those of chloropentaethoxysilico-ethane. It is actedon by alkali, with the evolution of hydrogen.The refractive indices recorded in this paper were kindly deter-mined by Mr. H. R. Nettleton, at the author's request.The author desires to express his thanks t o the Research FundCommittee of the Chemical Society for a grant which defrayed asmall part of the cost' of this investigation; also t o the Senate ofLondon University for a grant from the Dixon Fund whichdefrayed the main part of the expenses.~ ~ 1 I : K ~ E C K COLLROE,IAox~ox, E.C